R to @OmicronData: ...would you be surprised to hear this person meets the criteria for an autism diagnosis?
Even though this could equally be the description almost of a sullen, gaming, teenager?
R to @OmicronData: If I said there was a person who did the following:
- didn't make eye contact when talking to you
- had relatively few friends, and didn't seek to make them
- repeats certain words heavily
- doesnt usually start conversations
- had a particular special interest above all others
To those who think autism has increased significantly (by which I mean, the numbers affected have increased significantly), I ask - what do you think of when you think of autism?
I suspect that the perception is vastly different from the actual (and the diagnostic criteria).
Occam's razor applies here.
Definition of "autism" changed (substantially) in the early 1990s.
And awareness is hugely different now than 40 years ago.
Everything points to an increase in recognition and diagnosis, not an increase in actual prevalence.
Fear messaging and guilting parents is bad public health
R to @OmicronData: This very strongly demonstrates that the causal link is actually an undocumented confounder, *not* the painkiller itself.
And, as will come to a surprise to basically no-one who knows anything about this stuff, the observed risk ratio here is actually very similar to that observed when taking painkillers *other than* Tylenol:
https://pmc.ncbi.nlm.nih.gov/articles/PMC7170759/
R to @OmicronData: In other words, it comes as literally zero surprise to me that a study finds higher painkillers use in mothers who ultimately have autistic children (which is all the study actually finds).
But this is in no way the same as saying that use has *caused* this issue.
R to @OmicronData: This is your regular reminder that correlation ≠ causation.
R to @OmicronData: In other words, simply by virtue of the child having autism, the mother is roughly 5x more likely to have autism, and hence is more likely on average to have more chronic pain.
R to @OmicronData: I.e. by virtue of having an autistic child, you are predisposed, on average, to take more painkillers*anyway*, prior to having that child.
This would explain why sibling studies (i e. Where this uncontrolled confounder is effectively controlled), shows no such correlation.
R to @OmicronData: All 3 of these conditions, and others, cause, or can cause...physical pain.
In addition to this, while studies vary somewhat, there is a general consensus that there is at least a slightly heightened sensitivity to pain in autism generally.
R to @OmicronData: First of all, we know that autism is highly heritable - estimates range from 40-80%. Overall, this ultimately means that as an estimate, there is a roughly 5 times greater likelihood that mother has autism if a child has autism, than if a child doesn't have autism.
R to @OmicronData: We *also* know that autism almost always has some comorbid conditions. For example, it clusters heavily with ADHD, ODD and other conditions.
But, importantly, it also is often comorbid with hyper mobility, chronic pain, and mental health conditions.
Let's just talk about the apparent correlation between Tylenol and autism (and let's presume there is in fact a correlation for the sake of argument).
This is fully explainable by correlation rather than causation, when you actually account for what we know about autism.
One thing not remotely considered in these studies is that individuals with ASD are generally more sensitive to pain. And hence are likely to have greater painkiller use.
Any correlation between Tylenol and ASD could be solely explained by the hereditary nature of ASD.
Autism hasn’t suddenly skyrocketed. The numbers look bigger today because our definitions and awareness have changed.
In the 1940s, autism was first described as its own condition. Through the 50s and 60s it was often misdiagnosed as childhood schizophrenia. In the 80s it finally became its own diagnosis. In the 90s, the definition widened to include Asperger’s and related conditions. By 2013, everything was brought together under Autism Spectrum Disorder.
Today we know autism is a natural brain difference with genetic roots. More people are identified now not because autism is new, but because the medical community has gotten better at recognizing and diagnosing it.
That’s why what happened today is so damaging. Trump stood on a national stage and treated autism like a tragedy instead of a spectrum. He ignored the diversity of autistic lives. He recycled long-debunked myths about vaccines and medicine. He gave mothers a reason to blame themselves. He confused weak correlations with causes. He spread fear about treating common conditions during pregnancy. He erased autistic voices. And he politicized autism in a way that disrespected the people actually living it.
By doing this, he mocked decades of progress in understanding and acceptance for people with neurodivergence.
I am so frustrated by what I saw today. My heart aches for every advocate who has fought for awareness, dignity, and inclusion only to watch their work undermined in a single speech. You deserve so much more. We all do.
The announcement on autism was the greatest display of ignorance, unfounded blather, dangerous advice, recycled old nonsense and outright malpractice advice since Trump recommended bleach for Covid. A total laughingstock and embarrassment.
So confounders are not important to take into account in observational studies?
For example, consider the Swedish nationwide cohort study of acetaminophen and autism published in 2024 (https://pubmed.ncbi.nlm.nih.gov/38592388/), which contains 2.5m children born between 1995-2019 with followup through 2021. Below is a table of clinical characteristics of those children exposed to acetaminophen while in the womb or not.
Are you telling me that you would simply compare the raw autism rate between the two groups (0.7% vs. 0.4%) to assess whether acetaminophen was significantly associated with autism, and a potential causative factor, while ignoring the confounding factors?
Which are extensive, but that women whose children were exposed to acetaminophen in the womb had:
More infection (6.7% vs. 4.4%)
More chronic pain (1.7% vs. 0.8%)
More asthma (1.4% vs. 0.8%)
More rheumatoid arthritis (0.7% vs. 0.1%)
More migranes (0.4% vs. 0.2%)
More fever (0.4% vs. 0.2%)
Higher use of:
Opioids (15% vs. 2.4%)
Antimigrane medication (2.5% vs. 0.7%)
More antidepressants (5.7% vs. 3.1%)
More pscholeptic medications (4.0% vs. 1.0%)
More antiseizure medications (1.0% vs. 0.4%)
A higher % with >10 health care visits year before pregnancy (1.9% vs. 0.7%)
Do you REALLY think that someone who suggests that these confounders need to be taken into account has "conclusions shaped by politics not science", and the person who simply compares the raw rates, ignoring these massive imbalances, and declares acetaminophen use linked with autism is the one following science and not politics?
I sure hope that is not what you are saying, and not what you think.